Elements of episodic memory

Keen students of memory will recognise that the title of this post is an homage to the seminal book of the same title by the great memory researcher, Endel Tulving.  To my mind, Tulving’s Elements is one of the finest books that has been written about memory, along with William James’s Principles of Psychology and Dan Schacter’s Searching for Memory. (It’s quite possible that Charles Fernyhough’s forthcoming Pieces of Light may soon join that list).

In Tulving’s book, he describes how episodic memories of experienced events are unlikely to be stored as fixed, separate, discrete “memory traces”, but rather as “bundles” of features.  It makes sense, given the enormous number of events we may have to remember over a lifetime, that our brains would have evolved a more efficient strategy than simply storing each event separately, as a bound trace comprising all its different components.  The redundancy would be huge.  Instead, it appears that we store single representations of features distributed around the brain which are then shared between different event memories via associative networks. Tulving acknowledges that “we have no idea about the number and identity of features that the human mind or its memory system has at its disposal” (p. 161).  However, he speculates that “the features of the mind correspond to discriminable differences in our perceptual environment and to the categories and the concepts that the language we use imposes on the world.” (ibid)

Tulving’s conception of memories as comprising associations between visual, auditory, and other sensory features of events, as well as representations of thoughts and feelings we may have had when experiencing them, has been hugely influential.  Among the many reasons for this has been its ability to explain aspects of forgetting.  One of the principal causes of forgetting from episodic memory appears to be similarity between the features comprising different events.  Our memories are very susceptible to interference, either from previously-encoded events (so-called “proactive” interference) or from subsequent events disrupting earlier memories (“retroactive” interference).  The extent to which this interference leads us to confuse different events depends, in large part, on how much the features of those events overlap, or are similar.

A simple but striking way in which this can be demonstrated is the “release from proactive interference” effect, described by Delos Wickens.  If participants are asked to remember a string of consonants, their memory diminishes as the list grows longer because of interference between the items.  However, if the task involves remembering consonants for the first few trials and then switches to numbers, memory performance reverts to almost perfect levels.

The release from proactive interference effect

The release from proactive interference effect can be generalised to many other sensory features of events, and even to more abstract features such as meaning.  For example, a similar effect is observed if the first few trials comprise vegetables and then the stimuli switch to fruit (e.g., “spinach”, “beans”, “potato”, “orange”, ...)  A more recent experiment even found that the effect could be generalised to TV news bulletins.  Gunter and colleagues asked participants to watch a series of TV news items before asking them to recall the content of the stories they had just watched.  If the news stories all related to home news or all to foreign news, the typical effect of proactive interference was observed.  However, if the fourth news item was a different topic to the others (home news followed by world affairs, for example), release from proactive interference occurred (shown by the dashed line in the figure to the right).

Despite the prevalence of such interference effects, typically we’re actually quite good at discriminating similar events from one another.  Even when tasks are specifically designed to manipulate the extent to which the features of different events may overlap, people are often able to remember each experience pretty well, without getting them confused.

It is this fascinating ability that we are exploring with our online memory experiment, in collaboration with the Guardian, which we have launched this week.  We are hoping that thousands of people will take part, and make this the biggest memory experiment ever.

Anybody can participate by clicking http://www.guardian.co.uk/memorystudy to go to the Guardian experiment website and test their memory abilities for free from the comfort of their own homes.

Please do take part!

When the data are in, I’ll report back on what exactly the experiment was about, and what we found.

Gunter, B., Berry, C., & Clifford, B. (1981). Proactive interference effects with television news items: Further evidence. Journal of Experimental Psychology: Human Learning & Memory, 7 (6), 480-487 DOI: 10.1037/0278-7393.7.6.480

Wickens, D., Born, D., & Allen, C. (1963). Proactive inhibition and item similarity in short-term memory Journal of Verbal Learning and Verbal Behavior, 2 (5-6), 440-445 DOI: 10.1016/S0022-5371(63)80045-6